The period of adolescence involves dramatic changes in adolescents’ social networks, and entails extended reliance on friends and peers as well as investment in close mature intimate non-romantic friendships with best friends (Allen 2008; Way and Greene 2006; Selfhout et al. 2009). These intimate relationships are characterized by tenderness, self-disclosure, support, mutual assistance, and openness to sharing thoughts and feelings without fear of losing the self in the relationship (Montgomery 2005; Selfhout et al. 2009). With maturation, adolescents incorporate aspects of trust, loyalty, commitment, and exclusivity which sets the stage for future mature romantic relationships (Berndt 2004; Demir and Urberg 2004). Although in early adolescence the interest in close intimate relationships is mostly directed toward partners of the same sex, in middle and in late adolescence it is increasingly directed toward partners of the opposite sex (Steinberg 2008; Zimmer-Gembeck et al. 2012).

Developing an intimate relationship with a best friend has a significant influence on adolescents’ wellbeing. Studies suggest that a close friendship is positively associated with better psychological adjustment (Chou 2000; Zarabatany et al. 2004; La Greca and Harrison 2005). Adolescents who have no close friends were found to be at greater risk experience low self-esteem, loneliness, anxiety, and depression (Güroğlu et al. 2007; La Greca and Harrison 2005; Selfhout et al. 2009).

Theories of relational dynamics (Boszormenyi-Nagy and Krasner 1986), psychodynamic thinking (Winnicott 1965), developmental perspectives (Bowlby 1980), and family systems theories (Minuchin 1974) all argue that child development and the shaping of the family system are ongoing inter-dependent processes in which the psychological configuration of the child is part of the organization of the family. Researchers have emphasized the importance of clear boundaries between parent and child, in which the parent affords most of the scaffolding and support for the development of child’s functioning, thus enabling children to meet their emotional needs (Byng‐Hall 2008; Kerig 2005).

Normative transitions during adolescence are accompanied by alterations in parent-adolescent dynamics, thus requiring both to reconstruct the nature and the boundaries of their relationships by negotiating themes of connectedness and separatedness (Bush 2000). Various developmental researchers have underlined the role of clear parent-adolescent psychological boundaries for the formation of a sense of connectedness (Conger et al. 2001; Donnellan et al. 2005; Rubin et al. 2004). By contrast, deficiencies in parent-adolescent psychological boundaries may hinder adolescents’ ability to feel a sense of autonomy or connectedness (Kerig 2005; Rowa et al. 2001).

One particularly severe breach of parent-adolescent psychological boundaries is known as parentification (also termed role reversal). The term refers to an interaction in which the parent turns to the child for nurturance and support. This can involve practical and/or emotional role reversal in which adolescents must abandon their own needs for validation, guidance, and security to fulfill their parents’ self-absorbed needs (Earley and Cushway, 2002; Kerig 2005; Minuchin 1974).

Researchers differentiate between adaptive and destructive parentification. Adaptive parentification is relatively moderate in its intensity and is supported by the extended family. It aligns with cultural norms and the developmental phase of the adolescent, or is temporary in nature and allows for the development of empathy, self-competency and social skills resulting from the child’s contribution to the cohesion of the family (Byng‐Hall 2008). The parental role is openly delegated by the parents in the presence of siblings and the caring activities are supported, validated and are perceived as fair (Byng-Hall 2002; 2008; Saha 2016). By contrast, destructive parentification involves exaggerated emotional caregiving and instrumental responsibility. This forced obligation does not coincide with the adolescent’s development stage and interferes with adolescent’s individuation and identity formation (Burton 2007; Byng-Hall 2002, 2008; Kerig 2005; Jurkovic 1997). This situation may derive from the absence of a parent, dysfunction of a parent caused by mental illness, substance use, disability, immigration, parental conflict or divorce, parents with insecure ambivalent and disorganized attachment styles, or a trans-generational role-reversed family script and strategies (Fitzgerald et al. 2008; Macfie et al. 2005; Oznobishin and Kurman 2009; Ponizovsky et al. 2012). In this case, the parents are unable or unwilling to give the adolescent the required care, probably because they themselves need support and security (Boszormenyi-Nagy and Krasner 1986).

The findings on the impact of parentification on adolescents depict impairment in adolescents’ intimacy and socio-emotional functioning. For instance, parental role-taking, including asking for input on decisions or seeking advice on adult matters was shown to predict externalized and internalized behaviors among high risk American mother-adolescent dyads (Khafi et al. 2014) as well as substance use, early sexual behavior, and conduct problems in samples of families affected by maternal HIV (Lester et al. 2010; Stein et al. 1999). Internalizing behaviors were also found among European samples of adolescents living with a parent (either mother of father) suffering from mental health problems (Van Loon et al. 2017). Greater maternal or paternal disability from illness was related to higher caregiving in young people, which in turn was associated with higher adolescent distress, lower positive affect (Pakenham et al. 2006; Pakenham and Cox 2012), and health problems (Ireland and Pakenham 2010).

Enmeshment (also termed blurring of psychological boundaries) refers to a type of insufficient parent-child boundary that involves a lack of acknowledgement of the differentiation between parent and adolescent such that the adolescent is perceived as an extension of the parent (Kerig 2005; Werner et al. 2001). When this boundary is blurred, individuation and socio-emotional functioning are hindered, resulting in the adolescent’s inability to form a solid sense of self or preserve clear emotional boundaries between the self and others (Kerig 2005). Studies have found associations between mother- or father-adolescent enmeshment and poor intimacy and social adjustment. These were reflected, for example, in internalization problems among American adolescents at a residential treatment facility (Jewell and Stark 2003). Nurturance pursuit were found among British (Manzi et al. 2006) and Israeli adolescents with enmeshed relationships with both parents (Mayseless and Scharf 2009).

Desires for acceptance and belonging, and the drive to avoid rejection are widely recognized human needs (Baumeister and Leary 1995). However, individuals differ in their disposition to perceive and react to rejection. Greater rejection can be triggered by a cognitive-affective inclination to anxiously expect, readily perceive, and intensely response to rejection cues which is known as ‘rejection sensitivity’ (Downey and Feldman 1996; Feldman and Downey 1994). Early attachment interactions involving rejection by caregivers may result in increased anxiety and anticipation of rejection in future relationships (Feldman and Downey 1994; Romero-Canyas et al. 2010) and make individuals hyper-attentive to signs of rejection (Romero-Canyas et al. 2010). When they encounter rejection cues, however negligible or vague, they readily perceive intended rejection, suspicion, and feel rejected. These feelings can promote exaggerated behavioral and emotional reactions including anger as expressed in physical, verbal and nonverbal hostility as well as anxious reactions manifested in self-silencing, passive aggressiveness, and social withdrawal (Ayduk et al. 1999; Downey et al. 2000). This defense mechanism against future rejection may undermine the likelihood of maintaining close intimate relationships, and ultimately lead to future rejection and alienation (Ayduk et al. 2003; Downey et al. 2004).

Numerous studies have examined the deleterious consequences of rejection sensitivity on social maladjustment among early to late adolescents. For example, anxious and angry rejection expectations have been shown to be associated with social withdrawal, social avoidance and distress when encountering new situations and peers in elementary school age children and early adolescents (Bowker et al. 2013; London et al. 2007). Rejection sensitivity was correlated with loneliness and social anxiety among early adolescents (Rowe et al. 2015; Rudolph and Zimmer‐Gembeck 2014). Similarly, rejection sensitivity was linked to a decrease in social competence over a three-year period in late adolescence (Marston et al. 2010). With respect to romantic relationships, adolescents who were sensitive to rejection reported higher levels of self-silencing behaviors within their romantic partners (Harper et al. 2006).

Parent-adolescent relationships can also contribute to rejection sensitivity. For instance, parents’ attachment security was negatively correlated with early adolescents’ rejection sensitivity (Natarajan et al. 2011). Rejection sensitivity among early adolescents was the highest when parental acceptance was low (McLachlan et al. 2010). Finally, heightened perceptions of parental rejection (McDonald et al. 2010; McLachlan et al. 2010; Rudolph and Zimmer‐Gembeck 2014) predicted greater rejection sensitivity, which in turn was associated with early adolescents’ greater feelings of loneliness (McDonald et al. 2010).

Growing evidence suggests that cultural values such as family reciprocity, role flexibility and connectedness may moderate the family process underlying the developmental effects of parentification and enmeshment. Compared to adolescents from European American families, adolescents from collective societies with African, Asian, and Latin American backgrounds are more likely to take on a caregiving role including both instrumental and emotional caregiving (Burton 2007; Fuligni et al. 1999; Phinney et al. 2000). Whereas the Jewish Israeli culture is historically communal and collectivistic in that it emphasizes the centrality of the family, and aspects of dependency have been found to characterize Israeli parent-child relationships (Goldenr 2015; Mayseless and Scharf 2003), Israeli society today tends to adhere to the North American individualistic model in which independency, autonomy and self-determination dominate (Mayseless and Scharf 2003).

Studies on parentification and enmeshment in Israeli adolescents are relatively rare and have mainly examined immigrants from the Former Soviet Union and concentrated on the impact of parentification on their internal distress. For instance, adopting a parental role was positively related to these immigrant adolescents’ emotional distress. Language brokering was negatively related to adolescents’ self-esteem and self-efficacy (Oznobishin and Kurman 2009). However, taking the role as parental consoler, confidante or advice-giver in immigrant adolescents (but not in Israeli-born adolescents) was associated with more positive coping with stressful life events (Walsh et al. 2006).

Given the extensive literature on the associations between mother-adolescent parentification and enmeshment with separation anxiety, internalizing problems and poor social adjustment, as well as the association between adolescents’ rejection sensitivity and deficiencies in adolescents’ ability to forge close relationships, the current study examined the contributions of parentification and enmeshment to adolescents’ rejection sensitivity and intimacy. It was hypothesized that higher levels of parentification and enmeshment with mothers and adolescents’ rejection sensitivity would be directly negatively correlated with adolescents’ intimacy. Furthermore, using a mediation structural equation model, our main hypothesis was that higher levels of parentification and enmeshment with mothers would positively contribute to adolescents’ rejection sensitivity, which in turn would contribute to a lower level of intimacy with a same-sex best friend (see Fig. 1).

Fig. 1
figure 1

The mediation model. *p < .05, **p < .01, ***p < .001, N = 334

Full size image

Method

Participants

Three hundred thirty-four (N = 334) early (age 10–13, 33%) and middle adolescents (age 14–16, 67%) took part in the study utilizing a convenience sample. A power analysis using the Gpower computer program indicated that a total sample of 89 people would be needed to detect large effects (ƒ2 = .30) with 95% power using a linear multiple regression with an alpha of .05. The sample was composed of 8th and 9th graders drawn from three mid- to high- SES middle schools. Of the participants, 55% were girls and 45% were boys. The mean age of the adolescents was 13.95 (range 12.50–15.50; SD= .69) of whom 84% came from two-parent families, 15% from divorced families, and 1% from single parent families; 94% were born in Israel, and the others were immigrants (mostly from the Former Soviet Union). All participants spoke Hebrew.

Procedure

After receiving ethical approval from both the Ministry of Education and from the Committee to Evaluate Human Subject Research of the Faculty of Health Sciences and Social Welfare of the University of Haifa (#(938, informed consent was obtained from both the parents and the adolescents. A questionnaire booklet was administered in the school setting in the adolescents’ classrooms during a 45-min session. The second and the third authors introduced the project, read a few sample items out loud, and demonstrated how to complete the questionnaires. Participants were assured of the confidentiality of their responses. This work is an extension of a previous study that examined the associations between various types of parent-adolescent relationships and the construction of their true self (Goldner et al. 2017). LG: initiated the study, conducted the data analysis and write the final draft. AA and SCS: gathered the data and were involved in writing the article.

Measures

Intimacy with a same-sex best friend

The 32-item Intimate Friendship Scale (IFS; Sharabany 1994) was used to measure adolescents’ intimate friendships with their same-sex best friend and assessed frankness, sensitivity, attachment, exclusiveness, giving, obligation and commitment, taking part in shared activities, and loyalty. The adolescents were asked score the statements describing their relationship with their same-sex best friend and on a 7-point Likert-type scale ranging from strongly disagree (1) to strongly agree (7). The mean scores for all 32 items were then calculated to compute a total intimate friendship score. The scale was developed in Israel, has good psychometric properties (Sharabany 1994), and has been used on samples of various ages and cultures (e.g., Chou 2000; Oliva and Arranz 2005). The Cronbach alpha in the current study was .94.

Parentification and enmeshment

Adolescents completed the Parentification and the Enmeshment with the Mother subscales from the Inadequate Boundaries Questionnaire (IBQ; Mayseless and Scharf 2009). The parentification scale (five items) assesses the mother’s use of tactics of emotional and instrumental parentification. The enmeshment scale (five items) assesses the extent to which adolescents feel that their mothers perceive them as an extension of themselves. Items are rated on a 5-point Likert-type scale ranging from almost never (1) to almost always (5). The scale was developed in Israel and the internal reliabilities in the original study were consistently fair, ranging from .67 to .85. The Cronbach alphas in current study were .74 for parentification and .69 for enmeshment.

Rejection sensitivity

The Hebrew version of the Children’s Rejection Sensitivity Questionnaire (CRSQ; Downey et al. 1998) was used to measure adolescents’ anxious and angry expectations of rejection and their general sense of rejection from peers and teachers. The CRSQ describes 12 short hypothetical situations in which participants are asked to imagine requesting something from a significant other (i.e., peers or a teacher). Next, participants are asked to indicate on a 6-point Likert scale ranging from almost never (1) to almost always (6) how nervous (anxious expectations) and how mad (angry expectations) they would feel about the significant other not showing up. Finally, they are asked whether they thought the significant other would show up or not (expectation of rejection). The CRSQ has fair to good psychometric properties with Cronbach alphas ranging from .79 to .90 (Harper et al. 2006). The Cronbach alphas in the current study were .84 for angry expectations, .86 for anxious expectations, and .73 for expectation of rejection.

Data Analyses

The correlations between the adolescents’ background variables and the study variables were tested using t-test analyses and Pearson product-moment correlation coefficients. As can be seen in Table 1, this examination revealed a gender effect. By contrast, the t-test analyses did not reveal any difference in the variables across adolescents from single versus two-parent families or early versus mid-adolescents. Therefore, we examined the hypotheses for the entire sample by treating it as a single group while controlling for the gender effect.

Table 1 Differences in study variables according to gender
Full size table

To test the correlation between mother-adolescent parentification and enmeshment, rejection sensitivity and intimacy we used a series of partial Pearson product-moment correlation analyses between mother-adolescent parentification, enmeshment, rejection sensitivity and adolescents’ intimacy with best same-sex friend while controlling for gender, as presented in Table 2.

Table 2 Partial pearson product-moment correlations between the study variables controlled for gender
Full size table

To test the mediation model, Structural Equation Modeling (SEM) applying AMOS followed by bootstrap analyses (Preacher and Hayes 2008) was conducted. The maximum likelihood estimation method was used to obtain estimates of factor loadings, co-variances and residual variances. Several fit indices were used to evaluate the fit of the model. The Chi-square (χ²) test is a measure of exact fit. A significant χ²-value at an alpha-level of 0.05 indicates that the model does not fit the data (Browne and Cudeck 1992). A reasonable fit of the model is considered to be present if the χ2/df is approximately 3 or less. The CFI (The Comparative Fit Index) should vary between 0 and 1 and values of .90 or higher indicate an acceptable fit. For the Root Mean Square Error of Approximation (RMSEA), a value of .06 or lower is acceptable, for the Standardized Root Mean Square Residual (SMRS), values as high as .08 are deemed acceptable (Hu and Bentler 1999), and for the Goodness of Fit Index (GFI) a higher cutoff of .95 is more appropriate (Hooper et al. 2008). Coefficients of the direct and indirect effects are standardized, and thus values of 0.1, 0.3, and 0.5 can be interpreted as respectively ‘small’, ‘medium’, and ‘large’ effects (Cohen 1992). To build the model we used the two observed exogenous variables (parentification and enmeshment with the mother) as the predictors. These predictors were created from the original IBQ scales. One latent endogenous variable was used as a potential mediator and was comprised of the adolescents’ scores on angry, anxious, and rejection expectations. Finally, for the predicted variable, intimacy with the same-sex best friend was used as the observed variable. A regressed model was evaluated in which all the direct paths from the predictors (enmeshment and parentification with the mother) and the mediator (rejection sensitivity) to the predicted variable (intimacy with the same-sex best friend) were assessed. We included gender in the model as a control variable (for the inter-correlations of the model variables, see Table 2). The decision to input the relationship variables before rejection sensitivity stemmed from the theoretical justification that the early relationship with the mother impacts the tendency toward rejection sensitivity and its related consequences.

Results

Surprisingly, and as can be seen in Table 2, both mother-adolescent parentification and enmeshment were positively correlated with adolescents’ intimacy, whereas angry expectations and expectations of rejection were correlated negatively.

The mediation model is presented in Fig. 1, which contains the standardized estimates of the parameters in the structural model. The model exhibited a fairly good fit with the data (χ2 (9.86/6) = 1.64, p= .131, CFI = .99, NFI = .98, TLI = .98, RMSEA = .04, and SMRS = .03). Estimation of the mediation model demonstrated that mother-adolescent parentification was positively correlated with adolescents’ rejection sensitivity. This was negatively correlated with adolescents’ intimacy (Estimate = −.222, p = .008). In addition, the direct path between mother-adolescent parentification and adolescents’ intimacy was not significantly correlated. This indicates that adolescents’ rejection sensitivity fully mediated the correlation between mother-adolescent parentification and adolescents’ intimacy. Results from 500 bootstrap samples (which were all un-standardized) indicated that the 95% CI for the indirect effects did not include zero, thus indicating that the indirect effect was statistically significant. In particular, the bias-corrected bootstrap estimate of the indirect effect had a 95% confidence interval of −.125 to −.025. In contrast, rejection sensitivity did not mediate the link between mother-adolescent enmeshment and adolescents’ intimacy. Contrary to expectations, the direct path between mother-adolescent enmeshment and adolescents’ intimacy was significantly and positively correlated, implying a positive direct association between the two variables (Estimate = .210, p = .001).

Discussion

The current study was designed to shed light on the psychological mechanisms that contribute to the construction of intimacy with a same-sex best friend during early and mid-adolescence by exploring the contribution of mother-adolescent enmeshment and parentification as well as adolescents’ rejection sensitivity in a sample of early to mid-Israeli adolescents. Factors promoting adolescents’ intimacy have been examined in the context of attachment security and parental support (Dwyer et al. 2010); hence, the current examination may provide further insights into the nature of adolescents’ intimacy with their best friend. Note, however, that the cross-sectional data and the small effect sizes preclude generalization, and that it is important to replicate the finding in other samples as well.

With regard to parentification, the bivariate correlation indicated a positive correlation between mother-adolescent parentification and adolescents’ intimacy with a same-sex best friend. This positive association, which suggests the putative benefits of parentification on adolescents’ intimacy, may have to do with the low level of parentification in the current sample and may align with the notion of adaptive parentification as well as previous findings which have found an association between emotional caregiving in the family and pro-social behaviors among adolescents and young adults with a chronically ill/disabled family member (Ireland and Pakenham 2010). It is thus possible that in a situation with a low level of parentification, parentified non-clinical adolescents may exhibit their internalized caretaking role through interactions with others since these represent extensions of their childhood roles (Byng‐Hall 2008). By creating intimate relationships with their best friends, parentified adolescents may possibly obtain overt validation and recognition of their caretaking role and skills that they have acquired throughout their maturation process.

Nevertheless, when taking into account the contribution of adolescents’ rejection sensitivity as an intrapsychic mechanism that mediates the link between mother-adolescent’s parentification and intimacy with their same-sex best friend, the findings from the mediation model suggest that mother-adolescent parentification may be a risk factor for adolescent intimacy and become destructive. These findings are consistent with Van Loon et al.’s cross-sectional findings Van Loon et al. (2017) of an indirect effect of parentification on adolescents’ internalizing and externalizing problems via perceived stress. It is therefore conceivable that the struggle for attachment with their mothers while trying to satisfy their own emotional needs evokes feelings of rejection, which may generalize into difficulties in forging intimate relationships.

Contrary to expectations, both the bivariate correlations and the mediation model indicated that enmeshment with the mother positively predicted positively adolescents’ intimacy with their same-sex best friend. Thus, although patterns of enmeshment interfere with children’s development of an autonomous self, they may translate into positive aspects of connectedness, emotional involvement, and support-seeking which characterize both enmeshed and intimate relationships (Werner et al. 2001). Moreover, these positive associations may be contextual and cultural. There is evidence that the nature of maternal sensitivity across cultures varies in terms of the goals and beliefs underlying caregiving behaviors (van IJzendoorn 1990). Although maternal sensitivity in European-Western societies is based on values of agency, exploration, and independence in addition to aspects of availability, warmth, and responsiveness, sensitive mothers in collectivistic societies endorse values of over-involvement, dependency, and control (Keller et al. 2005; Takahashi et al. 2002). In particular, these mothers favor interpersonal fusion between caregiver and child by using physical contact and controlling behaviors (Bornstein 2012; Melendez 2005; Rothbaum et al. 2007). Since Israeli society is relatively collectivistic and highlights over-involvement among its members and family interconnectedness (Goldner 2015; Scharf and Mayseless 2010), the enmeshed relationship with mothers may be perceived as an expression of closeness and involvement which contribute to intimacy, rather than intrusiveness and invalidation of the self.

Several limitations of this study should be acknowledged. This study implemented a cross-sectional study design that was analyzed through Pearson product-moment correlations, which thus precludes drawing causal relationships. Future studies would benefit from using a longitudinal design to probe causality. Second, the effect sizes of the correlations were rather moderate, indicating that other variables such as the type and duration of parentification, adolescents’ perceptions of the fairness of parentification and the perceived enmeshment, mothers’ and adolescents’ mental state, and familial context could contribute to accounting for the results. Future studies should evaluate the model in light of these variables. Third, the present study was limited to adolescents’ self-report measures. Future studies would gain from including data from multiple respondents such as parents to reduce self-report biases. Fourth, the current study used one measure to assess both emotional and instrumental parentification, such that future studies may wish to make use of additional more fine-grained assessments to measure parentification. Furthermore, the data were collected solely on mother-adolescent parentification. Future studies could consider adding father-adolescent parentification. Finally, the results of this study reflect the reports of a non-clinical Israeli sample. Examining the hypotheses with diverse samples is called for before generalizing the results to other populations.

We declare that this work was carried out in the absence of any personal, professional, commercial, or financial relationships that could be construed as a potential conflict of interest.

Nevertheless, theoretically, the mediation model may lead to a more comprehensive scientific and clinical perspective on the impact of rejection sensitivity and parentification in the construction of adolescents’ intimacy, and underscores the relevance of psychoanalytic interpretations regarding separation and parentification (e.g., Winnicott 1965) to present-day developmental research. Furthermore, the direct findings between enmeshment and parentification and adolescents’ intimacy with a same-sex best friend support a multidimensional view of parentification and enmeshment as culturally implanted phenomena whose effects can be understood in terms of the cultural context in which they unfold.